Skip to content

Latest commit

 

History

History
403 lines (287 loc) · 10.7 KB

README.md

File metadata and controls

403 lines (287 loc) · 10.7 KB
Raw

OS Conductor

Gitter Travis Coveralls Issues Docs

Web services for OpenSpending, responsible for:

  • user authentication, identity and access control
  • package upload, management, and status
  • package search of os-package-registry
  • upload to the S3 datastore

os-conductor uses Flask web framework.

Quick start for development

Clone the repo, install dependencies from pypi, and run the server. See the docs for more information.

The os-types node utility is used to perform fiscal modelling for the processed datapackage. To install, use npm:

$ npm install -g os-types

Tests

With a running ElasticSearch server available on localhost:9200:

$ pip install tox  # install tox
$ tox

Environmental Variables

OS Conductor requires environmental variables to be set, either in the local environment or in a .env file in the root directory.

# Required settings

# Base URL for the application, e.g. 'http://localhost' or 'https://openspending.org'
OS_BASE_URL=
# Address for the postgres instance, e.g. postgresql://postgres@db/postgres
OS_CONDUCTOR_ENGINE=
# Address for ElasticSearch instance
OS_ELASTICSEARCH_ADDRESS=

# OAuth credentials. See the OAuth Credentials section below for details.
OS_CONDUCTOR_SECRETS_0=
OS_CONDUCTOR_SECRETS_1=
OS_CONDUCTOR_SECRETS_2=
OS_CONDUCTOR_SECRETS_3=

# AWS S3 credentials
OS_ACCESS_KEY_ID=
OS_SECRET_ACCESS_KEY=
OS_S3_HOSTNAME=
OS_STORAGE_BUCKET_NAME=

# Optional settings

# Address for memcached server, e.g. http://cache:11211
OS_CONDUCTOR_CACHE=

# Address for the redis os-api-cache server, e.g. redis
OS_API_CACHE=

# If this env var exists, the entrypoint script will check whether ElasticSearch is healthy before allowing os-conductor to start.
OS_CHECK_ES_HEALTHY=

# If using the fake-s3 docker container for development, openspending/fakes3, add these settings:
USE_FAKE_S3=True
OS_S3_PORT=4567

# comma-separated list of colon-separated api key and user id to enable secret-based authentication.
# API keys are strong secrets we make up for each client.
# userids are the user id of the user seen in loca storage or URLs after a user logs in via Google OAuth.
OS_CLIENT_API_KEYS=...apikey...:...userid...,...apikey2...:...userid2...

OAuth Credentials

OS Conductor needs credentials for authentication and authorization tasks. Credential values is set on OS_CONDUCTOR_SECRETS_<n> env vars. We provide a python script to help generate these values in docker/secrets/generate-secrets/to_env_vars.py.

  1. Create a Google OAuth Credentials and retain the Client ID and Secret Key
  2. Paste the Client ID and Secret Key values in to google.key and google.secret.key files respectively within the generate-secrets directory.
  3. Run the python script:
$ cd docker/secrets/generate-secrets
$ python ./to_env_vars.py
  1. Copy the generated env var key/values into your local environment or .env file in the root directory.

Admin tools

Various admin tools are available in the /tools directory. Some tools require dependencies to be installed from /tools/requirements.txt.

remove_package.py

Remove a named package (or packages) from the ElasticSearch index, and hence from searches and discovery within OpenSpending. Removing a package from the index won't remove it from the AWS datastore.

API Endpoints

Conductor current ships with the following blueprints, and their API endpoints.

Get authorized upload URL(s)

/datastore/authorize

Method: POST

Query Parameters:

  • jwt - permission token (received from /user/authorize)

Headers:

  • Auth-Token - permission token (can be used instead of the jwt query parameter)

Body:

JSON content with the following structure:

{
    "metadata": {
        "owner": "<user-id-of-uploader>",
        "name": "<data-set-unique-id>"
    },
    "filedata": {
        "<relative-path-to-file-in-package-1>": {
            "length": 1234, // length in bytes of data
            "md5": "<md5-hash-of-the-data>",
            "type": "<content-type-of-the-data>",
            "name": "<file-name>"
        },
        "<relative-path-to-file-in-package-2>": {
            "length": 4321,
            "md5": "<md5-hash-of-the-data>",
            "type": "<content-type-of-the-data>",
            "name": "<file-name>"
        }
        ...
    }
}

owner must match the userid that is in the authentication token.

Get information regarding the datastore

/datastore/info

Method: GET

Query Parameters:

  • jwt - permission token (received from /user/authorize)

Headers:

  • Auth-Token - permission token (can be used instead of the jwt query parameter)

Returns:

JSON content with the following structure:

{
    "prefixes": [
        "https://datastore.openspending.org/123456789",
        ...
    ]
}

prefixes is the list of possible prefixes for an uploaded file for this user.

Load a datastore package into the OpenSpending DB

/package/upload

Method: POST

Query Parameters:

  • jwt - permission token (received from /user/authorize)
  • datapackage - URL of the Fiscal DataPackage to load

Check on the status of uploading a package to the OpenSpending DB

/package/status

Method: GET

Query Parameters:

  • datapackage - URL of the Fiscal DataPackage being loaded

Returns:

{
    "status": "<status-code>",
    "progress": 123,
    "error": "<error-message-if-applicable>"
}
  • status-code: one of the following:
    • queued: Waiting in queue for an available processor
    • initializing: Getting ready to load the package
    • loading-datapackage: Reading the Fiscal Data Package
    • validating-datapackage: Validating Data Package correctness
    • loading-resource: Loading Resource data
    • deleting-table: Clearing previous rows for this dataset from the database
    • creating-table: Preparing space for rows in the database
    • loading-data-ready: Starting to load rows to database
    • loading-data: Loading data into the database
    • creating-babbage-model: Converting the Data Package into an API model
    • saving-metadata: Saving package metadata
    • done: Done
    • fail: Failed
  • progress: # of records loaded so far

Wil return an HTTP 404 if the package is not being loaded right now.

Toggle or set a package's privacy setting

/package/publish

Method: POST

Query Parameters:

  • jwt - permission token (received from /user/authorize)
  • id - Unique identifier of the datapackage to modify
  • publish - Publishing status, either:
    • true: force publish,
    • false: force private,
    • toggle: toggle the state

Returns:

{
    "success": true,
    "published": true  // or false
}

Search for specific packages

/search/package

Method: GET

Query Parameters:

  • jwt - authentication token (received from /user/check)
  • q - match-all query string
  • package.title - filter by package title
  • package.author - filter by package author
  • package.description - filter by package description
  • package.regionCode - filter by package region code
  • package.countryCode - filter by package region code
  • package.packageCode - filter by package region code
  • size - number of results to return

All values for all parameters (except jwt) should be passed as JSON values.

Returns:

All packages that match the filter.

If authentication-token was provided, then private packages from the authenticated user will also be included. Otherwise, only public packages will be returned.

[
    {
        "id": "<package-unique-id>",
        "model": { ... }, // Babbage model
        "package": { .... }, // Original FDP
        "origin_url": "<url-to-the-datapackage.json>"
    }
]

Check an authentication token's validity

/user/check

Method: GET

Query Parameters:

  • jwt - authentication token
  • next - URL to redirect to when finished authentication

Returns:

If authenticated:

{
    "authenticated": true,
    "profile": {
        "id": "<user-id>",
        "name": "<user-name>",
        "email": "<user-email>",
        "avatar_url": "<url-for-user's-profile-photo>",
        "idhash": "<unique-id-of-the-user>",
        "username": "<user-selected-id>" // If user has a username
    }
}

If not:

{
    "authenticated": false,
    "providers": {
        "google": {
            "url": "<url-for-logging-in-with-the-Google-provider>"
        }
    }
}

When the authentication flow is finished, the caller will be redirected to the next URL with an extra query parameter jwt which contains the authentication token. The caller should cache this token for further interactions with the API.

Get permission for a service

/user/authorize

Method: GET

Query Parameters:

  • jwt - user token (received from /user/check)
  • service - the relevant service (e.g. os.datastore)

Returns:

{
    "token": "<token-for-the-relevant-service>"
    "userid": "<unique-id-of-the-user>",
    "permissions": {
        "permission-x": true,
        "permission-y": false
    },
    "service": "<relevant-service>"
}

Note: as of yet: the permissions property is still returned empty. Real permissions will be implemented soon.

Change the username

/user/update

Method: POST

Query Parameters:

  • jwt - authentication token (received from /user/check)
  • username - A new username for the user profile (this action is only allowed once)

Returns:

{
    "success": true,
    "error": "<error-message-if-applicable>"
}

Note: trying to update other user profile fields like email will fail silently and return

{
    "success": true
}

Receive authorization public key

/user/public-key

Method: GET

Returns:

The conductor's public key in PEM format.

Can be used by services to validate that the permission token is authentic.

Read authentication JS library

/user/lib

Method: GET

Returns:

Authentication Javascript library with Angular 1.x binding.